Maceira et al. Journal of Cardiovascular Magnetic Resonance 2013, 15:29 http://www.jcmr-online.com/content/15/1/29

RESEARCH Open Access Reference right atrial dimensions and volume estimation by steady state free precession cardiovascular magnetic resonance Alicia M Maceira1, Juan Cosín-Sales2, Michael Roughton3, Sanjay K Prasad4 and Dudley J Pennell4*

Abstract Background: Cardiovascular magnetic resonance (CMR) steady state free precession (SSFP) cine sequences with high temporal resolution and improved post-processing can accurately measure RA dimensions. We used this technique to define ranges for normal RA volumes and dimensions normalized, when necessary, to the influence of gender, body surface area (BSA) and age, and also to define the best 2D images-derived predictors of RA enlargement. Methods: For definition of normal ranges of RA volume we studied 120 healthy subjects (60 men, 60 women; 20 subjects per age decile from 20 to 80 years), after careful exclusion of cardiovascular abnormality. We also studied 120 patients (60 men, 60 women; age range 20 to 80 years) with a clinical indication for CMR in order to define the best 1D and 2D predictors of RA enlargement. Data were generated from SSFP cine CMR, with 3-dimensional modeling, including tracking of the atrioventricular ring motion and time-volume curves analysis. Results: In the group of healthy individuals, age influenced RA 2-chamber area and transverse diameter. Gender influenced most absolute RA dimensions and volume. Interestingly, right atrial volumes did not change with age and gender when indexed to body surface area. New CMR normal ranges for RA dimensions were modeled and displayed for clinical use with normalization for BSA and gender and display of parameter variation with age. Finally, the best 2D images-derived independent predictors of RA enlargement were indexed area and indexed longitudinal diameter in the 2-chamber view. Conclusion: Reference RA dimensions and predictors of RA enlargement are provided using state-of-the-art CMR techniques. Keywords: Magnetic resonance, Heart, Right atrial volume, Dimensions, Reference values

Background the other, so their combined enlargement may better ex- Right atrial (RA) enlargement may occur in numerous press more remarkable structural remodeling and, more conditions including congenital heart disease, acquired importantly, the combination of these conditions might valvular disease, pulmonary disorders, and . be a better prognostic indicator of AF recurrence than Remodeling of the RA has also been reported in patients either alone. Another potential clinical value of RA with paroxysmal atrial (AF) [1-3], and recent measurement, since RA size is at least partially deter- data have shown that both left atrial (LA) and RA re- mined by the same factors that affect diastolic right ven- modeling are equally associated with recurrence of AF tricular (RV) filling, is its ability to act as an early after cardioversion [4]. Interestingly, RA and LA remod- marker of RV dysfunction, which often precedes systolic eling may coexist because one predisposes the heart to dysfunction in a variety of conditions affecting the RV. Furthermore, it may provide significant prognostic infor- mation in patients with chronic systolic heart failure [5] * Correspondence: [email protected] 4Cardiovascular Magnetic Resonance Unit, Royal Brompton Hospital, Sydney and pulmonary hypertension [6,7]. Street, London SW3 6NP, United Kingdom Full list of author information is available at the end of the article

© 2013 Maceira et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Maceira et al. Journal of Cardiovascular Magnetic Resonance 2013, 15:29 Page 2 of 10 http://www.jcmr-online.com/content/15/1/29

Atrial remodeling is most accurately estimated by Royal Brompton Hospital, and all subjects gave written measuring atrial volume, but in current clinical practice informed consent. this is not routinely performed. Cardiovascular magnetic resonance (CMR) is the gold standard technique for CMR measurement of ventricular dimensions and function CMR was performed with 1.5 T scanners (Siemens with reference ranges established from the Steady State Sonata and Avanto) using front and back surface coils Free Precession (SSFP) technique [8,9], and this has also and retrospective ECG triggering for capture of the en- been reported for the left atrium [10]. Some studies on tire cardiac cycle including diastole. All CMR scans were RA reference values have been published [11-13], but performed by the same operator. SSFP end-expiratory RA dimensions have not been extensively studied with breath-hold cines were acquired in the two (left and CMR as well as the systematic analysis of the influences right chambers) and four chamber views, with subse- of age, gender and body surface area (BSA). Therefore, quent contiguous short-axis cines from the atrioven- the aim of this study was to establish SSFP based refer- tricular (AV) ring to the base of the atria. Slice thickness ence values in normal subjects for RA dimensions nor- was 5 mm with no gap between slices. The temporal malized for independent influences such as gender, body resolution was 21 ±1 ms. Sequence parameters included surface area and age when required. We also aimed to repetition time/echo time of 3.2/1.6 ms, in-plane pixel determine the best predictors of right size of 2.1 × 1.3 mm, flip angle 60º, and acquisition time among 1D and 2D parameters, and relate RA volume to of typically 18 heartbeats. RA diameters and areas, as these are easily obtained in the clinical setting. CMR analysis Analysis was performed with a personal computer and Methods semi-automated software (CMRtools, Cardiovascular Normals and patients Imaging Solutions, London, UK). In all subjects (healthy For definition of normal ranges of RA dimensions we controls and patients) RA maximum volume was mea- studied 120 subjects, with 10 men and 10 women in sured as well as maximum diameters and areas, mea- each of 6 age deciles from 20 to 80 years. This cohort of sured in the 42-chamber and right 2-chamber views. healthy subjects has been used for defining left and right Atrial volume analysis included 2 steps: First, delineation ventricular and LA reference dimensions. Their baseline of the atrial endocardial border in all planes in all car- characteristics have been published previously [8]. In diac phases. Second, the systolic descent and twist of the brief, all subjects were normotensives (hypertension de- was calculated from tracking of the valve fined as systolic blood pressure ≥ 140 mmHg and/or dia- motion on the long axis cines, and used to correct stolic blood pressure ≥ 90 mmHg), asymptomatic, with for increase in atrial volume due to AV ring descent no known risk factors or history of cardiac disease, and (Figure 1). We included the atrial appendage and ex- with normal physical examination and electrocardiogram cluded the cava veins. Other approaches are possible, (ECG). Height, weight, blood pressure, total cholesterol, but the arguments for and against are not decisive. All HDL and B-natriuretic peptide were measured in all. diameters and areas derived from 2D images were mea- BSA was calculated according to the Mosteller formula sured in the phase of the corresponding cine sequences [14]. The risk over 10 years was at which the atrial size and volume measurements were calculated [15]. BNP levels were 2.5 ±2.1 pg/mL (range at a maximum. The longitudinal diameter was measured 0.5 – 12.0), and all were in the normal range (<100 pg/ form the midpoint of the line between the lateral mL) [16]. Therefore, as far as it was possible to ascertain and septal (or superior and inferior in the 2-chamber with conventional noninvasive techniques, all the appar- view) insertion of the tricuspid valve to the roof of ently healthy volunteers had a normal cardiovascular the right atrium. Transverse diameter was measured per- system with no high blood pressure and no evidence of pendicular to the midpoint of the longitudinal diameter heart failure. In a second step, and in order to define the (Figure 2). best 1D and 2D predictors of RA enlargement, a group of 120 patients (60 men and 60 women, age range 20– Statistical analysis 80) that were referred to CMR for clinical reasons, who This was carried out with the statistical software SPSS Sta- were in sinus rhythm and who agreed to participate in tistics 17.0 (IBM, United States). All atrial parameters the study, were included. The main reasons for referral were found to satisfy a normal distribution using the to CMR have been published elsewhere [10] and are Kolmogorov-Smirnov test and summary data for these briefly summarized in Table 1. Research was in compli- variables are presented as mean ± SD. Intra and ance with the Helsinki Declaration. The study was ap- interobserver reproducibility were tested in 50 subjects proved by the institutional Ethics Committee of the belonging to the cohort of healthy subjects. Simple linear Maceira et al. Journal of Cardiovascular Magnetic Resonance 2013, 15:29 Page 3 of 10 http://www.jcmr-online.com/content/15/1/29

Table 1 Baseline characteristics of the healthy subjects and the patient group (mean ± SD) Healthy subjects Patients N 120 120 Males 50% 53% Age [yr] (min, max) 49 ± 17 (20, 80) 65 ± 12 (20, 80) Height [cm] 171 ± 9 163 ± 9 Weight [kg] 72 ± 13 76 ± 13 Body surface area [m2] 1.83 ± 0.18 1.82 ± 0.18 Body mass index [kg/m2] 24±4 29±5 Heart Rate [bpm] 66 ± 10 69 ± 13 Systolic blood pressure [mmHg] 124 ± 12 140 ± 25 Diastolic blood pressure [mmHg] 73 ± 7 77 ± 14 Pathology (n) Ischemic heart disease - 46 Coronary risk factors - 35 Hypertensive heart disease - 12 - 13 Dilated - 4 Restrictive cardiomyopathy - 2 Congenital heart disease - 3 - 2 Hypertrophic cardiomyopathy - 1 Arrhythmogenic right ventricular cardiomyopathy - 1 Pericardial disease - 1 regression was used to model the indexed data and to and 4% for longitudinal and transverse diameters in the construct reference ranges as mean and 95% confidence 4-chamber view, respectively. Interobserver variability intervals. Multivariable analysis was used to analyze varia- was 3.9% for RA volume, 5.2% and 5% for areas in the 2- tions in parameters due to age and gender. P values <0.05 chamber and 4-chamber views, 5.5% and 5.5% for longi- were considered significant. In the patient group, correla- tudinal and transverse diameters in the 2-chamber view, tions of 1D and 2D parameters with RA volume were and 5.8% and 5.1% for longitudinal and transverse diam- assessed with the Pearson’scoefficient.Logisticregression eters in the 4-chamber view, respectively. RA reference analysis was used to define the best predictors of RA en- values with differentiation into males, females and all largement among 1D and 2D parameters. Linear regres- subjects, without age breakdown, and sub-division into sion analysis was used to develop a method to predict RA absolute and body surface area normalized values are volume with these parameters. showninTable2,whichhaveapplicationinstudiesofun- sorted individuals. Parameters that showed differences with Results age are also depicted, with age breakdown, in Table 3. Baseline characteristics and summary results for the healthy subjects group Influence of body surface area on atrial parameters Table 1 summarizes the baseline characteristics of the BSA was significantly higher in males than in females healthy subjects included for defining normal reference (p <0.001). On multivariable analysis, BSA was found to values. The reproducibility study was undertaken by 2 have significant independent influence on all RA param- operators with more than 5 years’ experience in CMR. eters except on 2-chamber area, and on transverse diam- One operator assessed intraobserver variability and the eter in the 2-chamber view. other one, blinded to previous results, interobserver variability. Intraobserver variability [17] was 3.5% for RA Influence of age on atrial parameters volume, 3.6% and 1.8% for areas in the 2-chamber and No significant increase in RA volume with age was ob- 4-chamber views, 4.1% and 4.2% for longitudinal and served in either univariable or multivariable analysis. On transverse diameters in the 2-chamber view, and 4.1% univariable analysis there was a significant decrease in Maceira et al. Journal of Cardiovascular Magnetic Resonance 2013, 15:29 Page 4 of 10 http://www.jcmr-online.com/content/15/1/29

Figure 1 CMR analysis of atrial volumes. Right atrial endocardial borders were delineated in all planes in all cardiac phases with inclusion of the atrial appendage and exclusion of the cava veins. The systolic descent and twist of the tricuspid valve was calculated from tracking of the valve motion on the long axis cines. The phase at which the atrial volume is at a maximum was selected for quantification. absolute and normalized areas and transverse diameters ±SD 59 ±35 mL/m2). According to our own normal refer- (measured in the 2-chamber view) in males (p < 0.001), ence values reported in Table 2, 22 patients (11 males, 11 and in absolute and normalized areas in the 2-chamber females) had left atrial enlargement (RAVi >78 mL/m2 in view and normalized diameter in the 4-chamber view in males and >70 mL/m2 in females). We aimed to deter- females (p = 0.013, 0.046 and 0.043 respectively). On mine the best independent predictors of RA enlargement, multivariable analysis, age was an independent predictor for which multivariate logistic regression analysis with for- of absolute and normalized transverse diameters and ward selection procedure was performed for indexed 1D areas (measured in the 2-chamber view) (all p = 0.001). and 2D parameters. For the sake of simplicity, these pa- Variables with significant differences according to age rameters were included as categorical dichotomous var- are depicted in Table 3. iables (below or above each parameter’supperlimitof normal for all subjects). This analysis showed that the Influence of gender on atrial parameters best predictors of RA enlargement were indexed area All absolute right atrial volume, diameters and areas were and indexed longitudinal diameter in the 2-chamber significantly larger in males (all p < 0.05) except transverse (Table 4). diameter in the 2-chamber view. When these parameters were normalized to BSA, only longitudinal (4-chamber Estimation of right atrial volume from 2D based view), and transverse (2-chamber view) diameters showed dimensions in the patient group differences, being both higher in females. On multivariable All 1D and 2D parameters correlated significantly with analysis, gender had no significant independent influence RA volume. The best correlations were found for areas on any variable. measured in the 2-chamber (r = 0.904, p < 0.001) and 4-chamber views (r = 0.868, p < 0.001) (Table 5). Finally, Predictors of atrial enlargement in the patient group linear regression analysis was used in the patient group The baseline characteristics of the patient group are also in order to estimate RA volume (Table 6). All non- depicted in Table 1. In this group of patients RA volumes indexed 1D and 2D measurements were included in showed a significant dispersion and, consequently, RA vol- multiple linear regression analysis and the equation ume index (RAVi) ranged from 18 to 253 mL/m2 (mean obtained was: Maceira et al. Journal of Cardiovascular Magnetic Resonance 2013, 15:29 Page 5 of 10 http://www.jcmr-online.com/content/15/1/29

A B C

T L

T

L

Figure 2 Measurement of right atrial parameters. Areas and diameters were measured in the phase of the cardiac cycle at which the atrial size was at a maximum. The figure shows the 2-chamber (top) and 4-chamber (bottom) views in which measurements were done. In B), longitudinal diameter (L) is obtained from the posterior wall of the right atrium to the center of the tricuspid plane, and transverse diameter (T) is obtained perpendicular to the longitudinal diameter, at the mid level of the right atrium. In C) measured areas are shown for both views.

RA volume = 3.08*(2C area) +3.36*(4C area) -44.4 with or all subjects, and in age deciles, when appropriate, or all an R2 of 0.895 (Figure 3). ages, in order to have applicability for comparison with This method was then tested in the group of healthy any other future research data set. For all ages and gen- subjects, and compared to the area-length method [18]. ders, a volume of 139 mL (75 mL/m2) was obtained as the Correlation coefficient with real RA volume was r = upper limit of normality. With regard to areas, the upper 0.824 for our method and 0.649 for the area-length limits of normality were 30 cm2 (15 cm2/m2) in the four method. Both methods were shown to underestimate chamber view, and 29 cm2 (16 cm2/m2) in the 2-chamber. real volume, with a mean difference of 3.38 ± 13 mL for The upper limits of normality for diameters in the our method and 22.7 ± 19.8 mL for the area-length 4-chamber and 2-chamber views were longitudinal 6.6 cm method (Figure 4). (3.6 cm/m2) and 6.5 cm (3.5 cm/m2), and transverse 5.8 cm (3.2 cm/m2) and 5.7 cm (3.2 cm/m2), respectively. Discussion RA volume measurements with CMR have been vali- Measurement of RA volume is not routinely performed dated in the past using an excised heart cast model [21], clinically although its prognostic value has been shown in but there is little peer-reviewed validated literature on RA a number of conditions, such as chronic heart failure [5], reference dimensions to compare our data [12,22,23], with [19] of pulmonary arterial hypertension very different results that are partly due to differences in [20]. This current study provides normal reference ranges imaging sequences, imaging views, acquisition, analysis for RA dimensions using state-of-the-art CMR acquisition and characteristics of the patients included. Some authors techniques and analysis in a healthy moderately large have published reference ranges for RA dimensions with population, which has been very well characterized for the CMR but we have not found any other study in which all absence of heart failure or any cardiomyopathy. CMR is a 1D, 2D and 3D parameters were measured with CMR. gold standard clinical technique to measure cardiac vol- Anderson et al. [13], measured maximal RA area and umes and function, so these data have significant clinical depth in end-systole in the four-chamber view and utility. The tables of results include all RA 1D and 2D pa- obtained an upper limit of normality of 23.5 cm2 for area rameters and volume, and are divided into males/females and 55.6 mm for depth, with no significant differences Maceira et al. Journal of Cardiovascular Magnetic Resonance 2013, 15:29 Page 6 of 10 http://www.jcmr-online.com/content/15/1/29

Table 2 Healthy subjects- Right atrial summary data for all ages (mean ± SD, 95% confidence interval) All Males Females Volume [mL] SD 20 100 (61, 139) 109 (64, 154) 91 (58, 124) Volume/BSA [mL/m2] SD 10.3 54 (34, 75) 55 (33, 78) 53 (36, 70) Area – 4ch[cm2] SD 3.8 22 (15, 30) 24 (15, 33) 20 (15, 26) Area/BSA – 4 ch [cm2/m2] SD 1.8 12 (8, 15) 12 (8, 16) 12 (9, 15) Longitudinal diameter – 4 ch [cm] SD 0.58 5.5 (4.3, 6.6) 5.6 (4.6, 6.7) 5.3 (4.3, 6.4) Longitudinal diameter/BSA – 4 ch [cm/m2] SD 0.32 3.0 (2.4, 3.6) 2.9 (2.5, 3.3) 3.1 (2.6, 3.6) Transverse diameter – 4 ch [cm] SD 0.55 4.7 (3.7, 5.8) 5.0 (3.7, 6.4) 4.5 (3.6, 5.4) Transverse diameter/BSA – 4 ch [cm/m2] SD 0.3 2.6 (2.0, 3.2) 2.6 (1.9, 3.3) 2.6 (2.2, 3.1) Area – 2ch[cm2] SD 3.95* 22 (14, 29) 23 (14, 31) 21 (14, 27) Area/BSA – 2 ch [cm2/m2] SD 2.27 * 12 (7, 16) 12 (7, 16) 12 (8, 16) Longitudinal diameter – 2 ch [cm] SD 0.5 5.4 (4.4, 6.5) 5.7 (4.6, 6.8) 5.1 (4.1, 6.1) Longitudinal diameter/BSA – 2 ch [cm/m2] SD 0.3 2.9 (2.4, 3.5) 2.9 (2.3, 3.4) 3.0 (2.0, 4.0) Transverse diameter – 2 ch [cm] SD 0.7 * 4.3 (3.0, 5.7) 4.3 (3.1, 5.5) 4.4 (3.3, 5.5) Transverse diameter/BSA – 2 ch [cm/m2] SD 0.4 * 2.4 (1.5, 3.2) 2.2 (1.5, 2.9) 2.6 (1.5, 3.7) BSA – body surface area; 4ch– 4-chamber view; 2ch– 2-chamber view. * Significant differences (p < 0.05) among age groups on multivariate analysis. between men and women (4-chamber RA area 22.5 vs which is higher than our results. We think that this dif- 18.8 cm2, RA depth 52.7 vs 51.5 mm). For the sake of ference in the upper limit of normality is attributable to simplicity, these authors concluded that an area a different, more heterogeneous, subject population, as <24 cm2 and a depth <58 mm included the upper 95th both the imaging sequence, acquisition method- with percentile of the normal range both for the left and retrospective gating- and image analysis were similar to right atria and best separated cardiomyopathic from ours. In fact the mean RA volume in Sievers’sstudyis normal hearts, these values are slightly lower than ours. 101 mL with a standard deviation of 30.2 mL, while we Sievers et al. [11], measured RA volumes in 70 subjects found a very similar RA volume, of 100 mL, with a using the short axis method and published an upper lower standard deviation, 20 mL, and thus a narrower limit of normality of RA volume of 170.4 mL (89.2 mL/m2), normal range and lower upper limit.

Table 3 Healthy subjects- Right atrial parameters with significant differences with age on multivariate analysis (mean, 95% confidence interval) 20-29 years 30-39 years 40-49 years 50-59 years 60-69 years 70-79 years All subjects RA area-2 ch [cm2] SD 4.5 24 (16, 32) 23 (15, 31) 22 (14, 30) 21 (13, 29) 20 (12, 28) 19 (11, 27) RA area/BSA- 2 ch [cm2/m2] SD 2.5 13 (9, 18) 13 (8, 17) 12 (8, 17) 12 (7, 16) 11 (6, 15) 10 (6, 15) RA transverse diameter- 2 ch [cm] SD 0.6 5.1 (3.7, 6.4) 4.7 (3.4, 6.1) 4.4 (3.1, 5.8) 4.1 (2.8, 5.5) 3.8 (2.5, 5.2) 3.5 (2.1, 4.8) RA transverse diameter/BSA- 2 ch [cm/m2] SD 0.4 2.8 (2.1, 3.6) 2.7 (1.8, 3.5) 2.5 (1.6, 3.4) 2.3 (1.4, 3.2) 2.2 (1.3, 3.1) 2.0 (1.1, 2.9) Males RA area-2 ch [cm2] SD 4.5 25 (16, 33) 24 (15, 33) 23 (14, 32) 22 (14, 31) 21 (13, 30) 21 (12, 29) RA area/BSA- 2 ch [cm2/m2] SD 2.5 13 (8, 17) 12 (8, 17) 12 (7, 17) 11 (7, 16) 11 (6, 16) 10 (6, 15) RA transverse diameter- 2 ch [cm] SD 0.6 4.9 (3.7, 6.2) 4.7 (3.5, 6.0) 4.5 (3.3, 5.7) 4.2 (3.0, 5.4) 4.0 (2.8, 5.2) 3.7 (2.5, 4.9) RA transverse diameter/BSA- 2 ch [cm/m2] SD 0.4 2.5 (1.8, 3.2) 2.4 (1.7, 3.1) 2.3 (1.6, 3.0) 2.1 (1.4, 2.8) 2.0 (1.3, 2.7) 1.9 (1.2, 2.6) Females RA area-2 ch [cm2] SD 3.4 24 (17, 31) 23 (16, 29) 21 (15, 28) 20 (13, 27) 19 (12, 25) 18 (11, 24) RA area/BSA- 2 ch view [cm2/m2] SD 2 14 (10, 18) 13 (9, 18) 13 (8, 17) 12 (8, 16) 11 (7, 15) 10 (6, 14) RA transverse diameter-2 ch [cm] SD 0.6 5.1 (4.0, 6.2) 4.8 (3.7, 5.9) 4.5 (3.4, 5.6) 4.2 (3.1, 5.3) 3.9 (2.8, 5.0) 3.6 (2.5, 4.7) RA transverse diameter/BSA- 2 ch [cm/m2] SD 0.4 3.1 (2.3, 3.9) 2.9 (1.8, 4.0) 2.7 (1.6, 3.8) 2.5 (1.4, 3.6) 2.3 (1.2, 3.4) 2.1 (1.0, 3.2) LA – left atrium; AP – anteroposterior; 2ch – 2-chamber view; RA – right atrium; BSA – body surface area; SD – standard deviation. Maceira et al. Journal of Cardiovascular Magnetic Resonance 2013, 15:29 Page 7 of 10 http://www.jcmr-online.com/content/15/1/29

Table 4 Predictors of right atrial enlargement according to RA volume index Univariate analysis: OR 95%CI p value Chi square 4-chamber longitudinal diameter (>6.6 cm) 13.6 6.97, 26.62 <0.001 NA 4-chamber transverse diameter (>5.8 cm) 5.0 2.41, 10.37 <0.001 18.7 4-chamber area (>30 cm2) 17.05 6.99, 41.17 <0.001 39.2 2-chamber longitudinal diameter (>6.5 cm) 9.6 4.14, 22.36 <0.001 27.7 2-chamber transverse diameter (>5.7 cm) 12.1 4.89, 30.12 <0.001 29 2-chamber area (>29 cm2) 21.6 6.83, 68.14 <0.001 27.4 4-chamber longitudinal diameter indexed (>3.6 cm/m2) 8.4 4.14, 17.12 <0.001 34.5 4-chamber transverse diameter indexed (>3.2 cm/m2) 8.4 3.29, 21.45 <0.001 19.8 4-chamber area indexed (>15 cm2/m2) 10.8 4.66, 25.18 <0.001 30.7 2-chamber longitudinal diameter indexed (>3.5 cm/m2) 12.0 3.71, 38.82 <0.001 17.2 2-chamber transverse diameter indexed (>3.2 cm/m2) 8.7 3.67, 20.44 <0.001 24.3 2-chamber area indexed (>16 cm2/m2) 38.5 9.55, 154.82 <0.001 26.4 Multivariate analysis (indexed parameters): OR 95%CI p value Chi square 2-chamber area indexed (>16 cm2/m2) 6.71 2.35, 19.13 <0.001 32.9 2-chamber longitudinal diameter indexed (>3.5 cm/m2) 7.78 2.14, 28.35 0.002 The upper limit of normal is shown in brackets according to Table 2.

We observed that nearly all non-indexed 1D, 2D and et al. [11], observed no age related differences in RA vol- 3D parameters were significantly higher in males, while ume with CMR. Aune et al. [24], measured RA volume these differences disappeared in most indexed parameters. by 3D echocardiography in 166 healthy subjects and This is in accordance with the findings by Sievers et al. found that normal aging does not increase RA size. On [11], who observed higher absolute volumes in males but the other hand Grapsa et al. [20], studied 62 consecutive no differences after adjusting by body surface area. patients with pulmonary arterial hypertension and ob- With respect to age, we found no differences in RA served increased RA sphericity index, which was a good volume with increasing age in this group of healthy indi- predictor of clinical outcome. viduals, and only found significant differences in both absolute and indexed transverse diameter and area in Comparison with echocardiographic studies and other the 2-chamber view. The RA is exposed to right ven- imaging techniques tricular diastolic pressure and, because of its thin walls, CMR does not require geometric assumptions to measure tends to dilate when pressure increases. In a healthy atrial volume, so volumes obtained with retrospectively population this is not the case and, though with increas- gated CMR are likely to differ significantly from those ing age a mild degree of myocardial stiffness could be obtained with 1D and 2D echocardiography. Echocardio- expected, no significant effect on RA volume was ob- graphic reference values have been quoted as 4.2 ± 0.4 cm served. Diastolic function parameters derived from ven- for RA depth and 14.0 ± 1.5 cm2 for RA area [25], lower tricular time-volume curves in this healthy population than our measurements. Differences are also related to the have been published elsewhere. This has been corrobo- greater precision of CMR compared with echo, the im- rated in a post-mortem study [19], and in a number of proved spatial resolution of endocardial border and in-vivo studies with different imaging techniques. Sievers slightly different anatomic views. Similarly Wang et al. [26], estimated with RA volume with the echocardio- Table 5 Correlations of 1D and 2D parameters with RA graphic area length method from the apical four-chamber volume (non-indexed parameters) view and obtained an upper limit of normality of 31 mL/ ’ Univariate analysis: Pearson s coefficient p value m2, far below our values, which can be explained by the 4-chamber longitudinal diameter 0.840 <0.001 different methodological approach and technical equip- 4-chamber transverse diameter 0.654 <0.001 ment. Whitlock et al. [27] compared RA volume estimated 4-chamber area 0.868 <0.001 using the echocardiographic area-length method and 2-chamber longitudinal diameter 0.746 <0.001 CMR and found that echocardiography caused a signifi- cant underestimation of RA volume. Currently, 3D echo is 2-chamber transverse diameter 0.773 <0.001 a more reproducible and robust method for measuring 2-chamber area 0.904 <0.001 RA volume. Aune et al. [24], obtained with 3D echo an Maceira et al. Journal of Cardiovascular Magnetic Resonance 2013, 15:29 Page 8 of 10 http://www.jcmr-online.com/content/15/1/29

Table 6 Predictors of right atrial volumes Univariate analysis: Coeff 95%CI p value r squared 4-chamber longitudinal diameter (cm) 48.4 42.7,54.1 <0.001 0.703 4-chamber transverse diameter (cm) 49.0 38.7,59.3 <0.001 0.428 4-chamber area (cm2) 6.72 6.0, 7.4 <0.001 0.752 2-chamber longitudinal diameter (cm) 42.7 35.1,50.3 <0.001 0.556 2-chamber transverse diameter (cm) 31.9 26.5,37.4 <0.001 0.573 2-chamber area (cm2) 4.6 4.2, 5.1 <0.001 0.806 4-chamber longitudinal diameter indexed (cm/m2) 63.7 52.2, 75.1 <0.001 0.508 4-chamber transverse diameter indexed (cm/m2) 77.1 58.6, 95.7 <0.001 0.365 4-chamber area indexed (cm2/m2) 11.1 9.7, 12.4 <0.001 0.698 2-chamber longitudinal diameter indexed (cm/m2) 66.4 52.4, 80.4 <0.001 0.470 2-chamber transverse diameter indexed (cm/m2) 51.4 41.4, 61.3 <0.001 0.512 2-chamber area indexed (cm2/m2) 7.8 7.0, 8.6 <0.001 0.772 Multivariate analysis: non-indexed measurements Coeff 95%CI p value r squared 2-chamber area (cm2) 3.1 2.6, 3.6 <0.001 0.895 4-chamber area (cm2) 3.4 2.6, 4.1 <0.001 Constant −44.4 upper normal value of 47 mL/m2 for the entire group, a significant underestimation of 12.06 mL, and worse cor- with a higher upper reference value for males (50 mL/m2) relation for 2D echo using single 4-chamber summation than females (41 mL/m2), still lower than our results. of disks algorithm (r = 0.79). This underestimation could Noteworthy, in this 3D echo study RA volume was found be due to a number of reasons including the higher spatial to be 15% higher than normal left atrial volume, similar to resolution of CMR, which permits more accurate border our findings comparing to our previous paper on LA vol- detection and better delineation of volumes within the tra- umes, and no significant correlation was found between beculae, low lateral resolution of the ultrasound beam, the RA volume and age. Keller et al. [23], validated echo de- gain dependent nature of the boundary echoes, and the rived RA volumes against CMR and found an excellent lower temporal resolution of 3D echo and reconstruction correlation for 3D echo derived RA volume (r = 0.91), with algorithms. These authors also suggested that CMR may overestimate RA volume by including the cava venous confluence, the appendage volume and the annular plane, 500 in our study we included the atrial appendage but care- y = 0,9098x + 9,9302 fully excluded the cava veins, and as for the tricuspid

2 400 r = 0.895 annular plane this was carefully delineated in the end- systolic phase. Cardiac computed tomography (CCT) has also been 300 used to measure RA volume, with reference values higher than ours. Lin et al. [28], measured RA volume with 64-row CCT in 103 healthy normotensive non- 200 obese volunteers and obtained a reference value of 111.9 ±29 mL with a reference range of 54.9-168.9 mL. This difference compared to our results could be at least in 100 part explained by differences in the recruited subjects, as RA volume, mL (estimated) this was not a population study, 57% of subjects were male and slightly older. On the contrary, Takahashi [29] 0 measured atrial volume with 320-slice computed tomog- 0 100 200 300 400 500 raphy and semi-automated 3 dimensional segmentation RA volume, mL (real) technique and found a normal value of 82.1 ± 44.1 mL, Figure 3 Scatterplot showing the correlation between real RA which is smaller than our results, though in this study volume and volume estimated with our method. The dotted line only 22 subjects were included and thus it is difficult to is the line of identity and the solid line is the line of regression. compare with our study. Maceira et al. Journal of Cardiovascular Magnetic Resonance 2013, 15:29 Page 9 of 10 http://www.jcmr-online.com/content/15/1/29

estimated volumes derived from this method and from A New method the traditionally used single plane area-length method. Both correlated well but caused a significant underestima- tion of RA volume, with worse accuracy for the area- length method (mean difference of 22.7 ±19.8 mL). Some studies have also compared methods of volume estimation with real volumes, with different results. Sievers et al. [11], also compared the single plane area-length method with the short axis method in 70 healthy subjects and found that the former overestimated RA volume.

Conclusions Difference in RA volume, mL in Difference Right atrial dimensions do not vary with gender after adjustment for body surface area and only a few show differences with age. References ranges are supplied with this report in both tabular and graphical form and are of RA volume, mL significant clinical and research utility for the interpret- B ation of CMR studies. Also, best predictors of RA en- Area-length method largement are provided.

Abbreviations

(RA): Right atrial; (CMR): Cardiovascular magnetic resonance; (SSFP): Steady state free precession; (AV): Atrioventricular. Professor Pennell is a consultant to Siemens and a director of Cardiovascular Imaging Solutions. The other authors have no conflicts to declare. This research was supported by CORDA and the British Heart Foundation. Research support was also received from Siemens.

Competing interests The authors declare that they have no competing interests.

Difference in RA volume, mL in Difference Authors contributions AMM- Study design, subject scanning, data selection, data processing, writing manuscript. JCS- Subject recruitment, data selection, data processing. MR- Study design, statistical analysis. SKP- Study design, subject scanning. DJP- Study design, editing manuscript. All authors read and approved the RA volume, mL final manuscript. Figure 4 Bland-Altman plot. The continuous line represents the Acknowledgements mean (bias) and the dotted lines represent the limits of agreement This research was supported by the National Institute for Health Research for A) our method of RA volume estimation and B) the area-length Cardiovascular Biomedical Research Unit of Royal Brompton Hospital and method, tested in the cohort of healthy subjects. Imperial College. Support was also received from CORDA, the British Heart Foundation and Siemens.

Author details Predictors of RA enlargement and estimators of RA 1Cardiac Imaging Unit, ERESA, Valencia, Spain. 2Department of Cardiology, volume Hospital Arnau de Vilanova, Valencia, Spain. 3Medical Statistics Department, Measurement of RA volume is desirable but may be time- Royal Brompton Hospital, London, UK. 4Cardiovascular Magnetic Resonance Unit, Royal Brompton Hospital, Sydney Street, London SW3 6NP, United consuming for daily clinical practice. Therefore, 1D and Kingdom. 2D parameters might be a valuable tool to assess RA size. The best independent indicators of RA enlargement in Received: 4 October 2012 Accepted: 28 February 2013 Published: 8 April 2013 our study were an area >16 cm2/m2 and a longitudinal 2 diameter >3.5 cm/m in the 2-chamber view. In the study References 2 by Anderson et al. [13], a non-indexed RA area <24 cm 1. Sanfilippo AJ, Abascal VM, Sheehan M, Oertel LB, Harrigan P, Hughes RA, and depth <5.8 cm were the parameters that best distin- Weyman AE. Atrial enlargement as a consequence of atrial fibrillation: a prospective echocardiographic study. Circulation. 1990; 82:792–7. guished normal from abnormal atria. We have not found 2. Stiles MK, John B, Wong CX, Kuklik P, Brooks AG, Lau DH, Dimitri H, Roberts- any other study with which to compare our data. Thomson KC, Wilson L, De Sciscio P, Young GD, Sanders P. Paroxysmal lone With respect to RA volume estimators, we found that atrial fibrillation is associated with an abnormal atrial substrate: characterizing the “second factor”. J Am Coll Cardiol. 2009; 53:1182–91. the best method included measurement of area in the 2 3. Kalifa J, Jalife J, Zaitsev AV, Bagwe S, Warren M, Moreno J, Berenfeld O, and 4-chamber views. We correlated real volume with Nattel S. Intra-atrial pressure increases rate and organization of waves Maceira et al. Journal of Cardiovascular Magnetic Resonance 2013, 15:29 Page 10 of 10 http://www.jcmr-online.com/content/15/1/29

emanating from the superior pulmonary veins during atrial fibrillation. 23. Keller AM, Gopal AS, King DL. Left and right atrial volume by freehand Circulation. 2003; 108:668–71. three-dimensional echocardiography: in vivo validation using magnetic 4. Akutsu Y, Kaneko K, Kodama Y, Suyama J, Li HL, Hamazaki Y, Tanno K, resonance imaging. Eur J Echocardiogr. 2000; 1:55–65. Gokan T, Kobayashi Y. Association between left and right atrial 24. Aune E, Baekkevar M, Roislien J, Rodevand O, Otterstad JE. Normal remodeling with atrial fibrillation recurrence after pulmonary vein reference ranges for left and right atrial volumen indexes and ejection catheter ablation in patients with paroxysmal atrial fibrillation. A Pilot fractions obtained with real-time three-dimensional echocardiography. Study Circ Cardiovasc Imaging. 2011; 4:524–31. Eur J Echocardiogr. 2009; 10:738–44. 5. Sallach JA, Tang WHW, Borowski AG, Tong W, Porter T, Martin MG, Jasper 25. Cheitlin MD, Armstrong WF, Aurigemma GP, Beller GA, Bierman FZ, Davis JL, SE, Shrestha K, Troughton RW, Klein AL. Right atrial volume index in Douglas PS, Faxon DP, Gillam LD, Kimball TR, Kussmaul WG, Pearlman AS, chronic systolic heart failure and prognosis. J Am Coll Cardiol Img. 2009; Philbrick JT, Rakowski H, Thys DM, Antman EM, Smith SC Jr, Alpert JS, 2:257–34. Gregoratos G, Anderson JL, Hiratzka LF, Hunt SA, Fuster V, Jacobs A, 6. Raymond RJ, Hinderliter AL, Willis PW, Ralph D, Caldwell EJ, Williams W, Gibbons RJ, Russell RO. ACC/AHA/ASE 2003 guideline update for the Ettinger NA, Hill NS, Summer WR, de Boisblanc B, Schwartz T, Koch G, clinical application of echocardiography: summary article. Circulation. Clayton LM, Jöbsis MM, Crow JW, Long W. Echocardiographic predictors 2003; 108:1142–62. of adverse outcomes in primary pulmonary hypertension. J Am Coll 26. Wang Y, Gutman JM, Heilbron D, Wahr D, Schiller NB. Atrial volume in a Cardiol. 2002; 39:1214–9. normal adult population by two-dimensional echocardiography. Chest. 7. Bustamante-Labarta M, Perrone S, De La Fuente RL, Stutzbach P, De La Hoz 1984; 86:595–601. RP, Torino A, Favaloro R. Right atrial size and tricúspide regurgitation 27. Whitlock M, Garg A, Gelow J, Jacobson T, Broberg C. Comparison of left severity predict mortality or transplantation in primary pulmonary and right atrial volume by echocardiography versus cardiac magnetic hypertension. J Am Soc Echocardiogr. 2002; 15:1160–4. resonance imaging using the area-length method. Am J Cardiol. 2012; 8. Maceira AM, Prasad SK, Khan M, Pennell DJ. Normalized left ventricular 106:1345–50. systolic and diastolic function by steady state free precession 28. Lin FY, Devereux RB, Roman MJ, Meng J, Jow VM, Jacobs A, Weinsaft JW, cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2006; 8:417–26. Shaw LJ, Berman DS, Callister TQ, Min JK. Cardiac chamber volumes, 9. Maceira AM, Prasad SK, Khan M, Pennell DJ. Reference right ventricular function and mass as determined by 64-multidetector row computed systolic and diastolic function normalized to age, gender and body tomography: mean values among healthy adults free of hypertension surface area from steady-state free precession cardiovascular magnetic and obesity. JACC Cardiovasc Imaging. 2008; 1:782–6. resonance. Eur Heart J. 2006; 27:2879–88. 29. Takahashi A, Funabashi N, Kataoka A, Yajima R, Takahashi M, Uehara M, 10. Maceira AM, Cosín-Sales J, Roughton M, Prasad SK, Pennell DJ. Reference left Takaoka H, Saito M, Yamaguchi C, Lee K, Komuro I, Nomura F. Quantitative atrial dimensions and volumes by steady state free precession cardiovascular evaluation of right atrial volume and right atrial emptying fraction by magnetic resonance. J Cardiovasc Magn Reson. 2010; 12:65–75. 320-slice computed tomography compared with three-dimensional – 11. Sievers B, Addo M, Breuckmann F, Barkhausen J, Erbel R. Reference right echocardiography. Int J Cardiol. 2011; 146:96 9. atrial function determined by steady-state free precession cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2007; 9:807–14. doi:10.1186/1532-429X-15-29 12. Mohiaddin RH, Hasegawa M. Measurement of atrial volumes by magnetic Cite this article as: Maceira et al.: Reference right atrial dimensions and resonance imaging in healthy volunteers and in patients with volume estimation by steady state free precession cardiovascular . Eur Heart J. 1995; 16:106–11. magnetic resonance. Journal of Cardiovascular Magnetic Resonance 2013 13. Anderson JL, Horne BD, Pennell DJ. Atrial dimensions in health and left 15:29. ventricular disease using cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2005; 7:671–5. 14. Mosteller RD. Simplified calculation of body surface area. N Engl J Med. 1987; 317:1098. 15. Joint British Cardiac Society. British hyperlipidaemia association and British hypertension society recommendations on coronary prevention. Heart. 1998; 80(Suppl 2):S1–29. 16. Maisel AS, Krishnaswamy P, Nowak RM, McCord J, Hollander JE, Duc P, Omland T, Storrow AB, Abraham WT, Wu AH, Clopton P, Steg PG, Westheim A, Knudsen CW, Perez A, Kazanegra R, Herrmann HC, McCullough PA. Breathing not properly multinational study investigators. Rapid measurement of b-type natriuretic peptide in the emergency diagnosis of heart failure. N Engl J Med. 2002; 347:161–7. 17. Bland JM, Altman DG. Statistical methods of assessing agreement between methods of clinical measurements. Lancet. 1986; 1:307–10. 18. Sievers B, Kirchberg S, Addo M, Bakan A, Brandts B, Trappe HJ. Assessment of left atrial volumes in sinus rhythm and atrial fibrillation using the biplane area-length method and cardiovascular magnetic resonance imaging with TrueFISP. J Cardiovasc Magn Reson. 2004; 6:593–600. 19. Platonov PG, Mitrofanova LB, Orshanskaya V, Ho SY. Structural abnormalities in atrial walls are associated with presence and persistency of atrial Submit your next manuscript to BioMed Central fibrillation but not with age. J Am Coll Cardiol. 2011; 58:2225–32. and take full advantage of: 20. Grapsa J, Gibbs SR, Zimbarra Cabrita I, Watson GF, Pavlopoulos H, Dawson D, Gin-Sing W, Howard L, Nihoyannopoulos P. The association of clinical • Convenient online submission outcome in patients with pulmonary arterial hypertension: study with real-time three-dimensional echocardiography. Eur Heart J Cardiovasc • Thorough peer review – Imaging. 2012; 13:666 73. • No space constraints or color figure charges 21. Jarvinen V, Kupari M, Hekali P, Poutanen VP. Right atrial MR imaging studies of cadaveric atrial casts and comparison with right and left atrial • Immediate publication on acceptance volumes and function in healthy subjects. Radiology. 1994; 191:137–42. • Inclusion in PubMed, CAS, Scopus and Google Scholar 22. Matsuoka H, Hamada M, Honda T, Kobayashi T, Suzuki M, Ohtani T, Takezaki • Research which is freely available for redistribution M, Abe M, Fujiwara Y, Sumimoto T. Measurement of cardiac chamber volumes by cine magnetic resonance imaging. Angiology. 1993; 44:321–7. Submit your manuscript at www.biomedcentral.com/submit